HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Fan, R. Articles by Braciale, T. J. Search for Related Content PubMed PubMed Citation Articles by Fan, R. Articles by Braciale, T. J.

Recognition of a Sequestered Self Peptide by Influenza Virus-Specific CD8⁺ Cytolytic T Lymphocytes¹

Rong Fan^2,*,, Scott S. Tykodi^2,*,§ and Thomas J. Braciale^3,*,,

^* Beirne B. Carter Center for Immunology Research and Departments of Microbiology and Pathology, University of Virginia Health Sciences Center, Charlottesville, VA 22908; and ^§ Division of Biology and Biomedical Science, Immunology Program, Washington University School of Medicine, St. Louis, MO 63110

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
The Ag receptors on CD8⁺ CTL recognize foreign antigenic peptides associated with cell surface MHC class I molecules. Peptides derived from self proteins are also normally presented by MHC class I molecules. Here we report that an H-2K^d-restricted murine CD8⁺ CTL clone directed to an influenza hemagglutinin epitope can recognize a peptide derived from the murine mitochondrial aconitase enzyme in association with H-2K^d molecules. Surprisingly, this self peptide is not normally displayed on the cell surface associated with the restricting MHC class I molecule. Several lines of evidence suggest that this self peptide, although requiring association with the K^d molecule for CTL recognition, is not associated with this or other MHC class I allele under physiologic conditions in intact cells. Rather, it is sequestered in the cytoplasm associated with a carrier protein and is released only upon cell disruption. These results suggest a means of restricting the entry of self peptide into the class I pathway. In addition, this finding raises the possibility that self peptides sequestered within the cell can, after release from damaged cells, interact with MHC class I molecules on bystander cells and trigger autoimmune injury by virus-specific CTLs during viral infection.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
The Ag receptors on CD8⁺ CTL recognize foreign Ag as short peptides, typically 8–11 residues in length, associated with cell surface class I MHC/ß₂-microglobulin (ß₂m)⁴ heterodimers. Entry of protein Ag into the class I processing pathway occurs primarily via the cytosol and is dependent on the proteolytic function of the cytosolic proteasome complex and the TAP1/TAP2 peptide transporter located in the endoplasmic reticulum (ER) and cis-Golgi membrane. Nascent class I complexes rapidly acquire peptide in the ER. The ternary complex of heavy chain, ß₂m, and peptide is required for efficient transport to the cell surface. Unstable complexes lacking peptide (or ß₂m) are preferentially retained in the ER by chaperone class molecules. Entry of Ag into the class I pathway via the cytosol appears to be the major common pathway in the cell regardless of the targeting of native Ag to subcellular compartments such as the ER and secretory pathway, nucleus, or mitochondria. This pathway can account for the majority of MHC peptide ligands identified to date. Nevertheless, a subset of class I peptide epitopes may be generated by alternative means. In certain model systems, processing of endocytosed soluble Ag has been shown to result in the production of class I-restricted epitopes (1).

It is likely that other molecules in addition to the proteasome and TAP transporters are required for peptide production and intracellular peptide trafficking for class I MHC presentation. It has been speculated that heat shock proteins (hsp) could serve as intermediaries in the movement of peptides between proteasome, TAP, and MHC complexes (2). The cytosolic hsp70 and hsp90, and the ER resident gp96/grp94 can be purified from immunogenic tumor lines associated with peptides that can stimulate tumor-specific CTL responses (3, 4, 5). The mechanism of class I priming by the hsp-peptide complex is thought to occur following receptor-mediated uptake of the hsp molecule by macrophage or dendritic cells (6). The extension of the range of peptides bound to hsp70 and gp96 to include peptides related to defined class I epitopes from viral and other model Ags has intensified speculation on a role for these molecules in intracellular peptide trafficking in the class I processing pathway (7, 8, 9).

Although CTL recognize MHC molecules presenting peptides derived from foreign Ag, the majority of peptides occupying MHC class I molecules at the cell surface arise from self proteins. The constitutive expression of class I at the cell surface is dependent on the continuous generation of peptides from cellular proteins. Whether all self proteins are equally available for class I processing is presently unknown. Roles for individual self peptide ligands are becoming more clearly established. Self peptide plays an essential role in both positive and negative selection of CTL precursors in the thymus (10). Consequently, the diversity of self peptide will probably impact the repertoire of T cell Ag receptors available in the periphery. There is also emerging evidence that self peptides may be actively involved in regulating T cell function and survival in the periphery (11). In addition, NK cells are known to express cell surface receptors distinct from the Ag receptors of CTL that recognize class I MHC on target cells in an allele-specific manner and that result in inhibition of target cell lysis. This interaction has been shown to require properly assembled peptide/MHC complexes and may have selectivity for specific self peptides (12, 13). Taken together, there is reason to speculate that the diversity of self peptides presented by class I MHC may have consequences for both CTL and NK cell function.

Each step in the class I processing pathway has been investigated for its contribution to the selection of peptide ultimately displayed by MHC at the cell surface. Allele-specific consensus sequence motifs of individual class I molecules are defined by the presence of a conserved or closely related residue at a fixed position within most peptide ligands (14). Motifs severely limit the array of possible class I epitopes that can be derived from a foreign Ag or self protein. While most MHC peptide ligands contain the allele-specific motif, not all motif-bearing peptides bind to MHC molecules (15). Parameters besides allele-specific sequence constraints thought to influence epitope selection include peptide affinity for MHC (16), selectivity imparted by the TAP transporter (17), and specificity of the proteases involved in generating peptides for class I loading (18, 19).

In this report we identify the sequence of the naturally processed K^d-restricted epitope from the A/Japan/57 influenza hemagglutinin (HA) recognized by CTL with specificity previously mapped to the transmembrane domain of the HA molecule (20, 21). In carrying out these experiments, we were surprised to discover a cross-reactivity of one of our K^d-restricted HA specific CTL clones for a K^d motif containing self peptide derived from the mitochondrial aconitase protein. Characterization of this peptide demonstrated a novel peptide phenotype. Although the aconitase peptide is present at high titer within cells, and its recognition is K^d restricted, several lines of evidence indicate that this peptide is not associated with K^d or other class I MHC allele under physiologic conditions in intact cells. Rather, it is sequestered in the cytoplasm associated with a carrier protein and is released only upon cell disruption. The implications of this finding of a cross-reactive, sequestered, self peptide for the presentation of self peptide to MHC class I molecules and for the pathogenesis of autoimmune tissue injury during virus infection are discussed.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Mice

BALB/c (H-2^d), C57BL/6 (H-2^b), and C3H (H-2^k) mice were purchased from Taconic Farms (Germantown, NY). ß₂m knockout mice backcrossed into the BALB/c background and H-2^d haplotype were a gift from T. Hansen (St. Louis, MO) and were bred under specific pathogen-free conditions at the University of Virginia animal care facility (Charlottesville, VA).

Cell lines

The HTR cell line is a variant of the P815 (H-2^d) murine mastocytoma selected for high transfection efficiency (22). The BHA cell line was derived from HTR by transfection with the hemagglutinin gene from the A/Japan/57 strain influenza virus. HTR and BHA were gifts from M. J. Gething (Dallas, TX). HTR was maintained in DMEM supplemented with 10% FCS, 1% 200 mM glutamine, and penicillin/streptomycin. BHA was cultured in similar medium supplemented with G418 to 0.5 mg/ml. RMA-S (23) was a gift from V. Engelhard (Charlottesville, VA). RMA-S was maintained in RPMI 1640 supplemented with 10% FCS, 1% 200 mM glutamine, HEPES to 20 mM, 2-ME to 5 x 10^-5 M, and penicillin/streptomycin. The K^d transfectant of the RMA-S cell line (denoted RMA-S/K^d) was a gift from M. Bevan (Seattle, WA). RMA-S/K^d was maintained in similar medium as RMA-S supplemented with G418 to 0.4 mg/ml. The S49.35 (H-2^d) lymphoma line (24) was maintained in DMEM supplemented with 10% FCS, 1% 200 mM glutamine, and penicillin/streptomycin. The ß₂m-deficient R1.E cell line (25) was maintained in RPMI 1640 supplemented with 10% FCS, 1% 200 mM glutamine, and penicillin/streptomycin. The S49.35 and R1.E lines were gifts from T. Hansen (St. Louis, MO). The procedures used to establish and maintain the influenza-specific clonal T cell lines have been reported previously (26). Clones were passaged weekly in the presence of A/Japan/57-infected, irradiated (2000 rad), BALB/c splenocytes in IMDM supplemented with 10% FCS, 1% 200 mM glutamine, 2-ME to 5 x 10^-5 M, 15 U/ml recombinant human IL-2 (BioSource, Camarillo, CA), and penicillin/streptomycin. Clones were routinely used in assays 5–7 days poststimulation.

Peptide extract from whole cell homogenates

The HTR and BHA cell lines were grown in flasks on a rocker platform to a density of 2–3 x 10⁶ cells/ml. Cell were pelleted and washed three times in PBS, pH 7.3, frozen in liquid nitrogen, and stored at -80°C until use. Spleens were harvested from 10–20 mice and dispersed as a single-cell suspension in neutral PBS, and total numbers of mononuclear cells were estimated. Splenocytes were pelleted, frozen in liquid nitrogen, and stored at -80°C until use. Cell pellets were thawed on ice and resuspended in 4 ml of 0.1% trifluoroacetic acid/10⁹ cells. For some experiments a protease inhibitor mixture including PMSF (2 mM), iodoacetamide (100 µM), aprotinin (5 µg/ml), leupeptin (10 µg/ml), pepstatin A (10 µg/ml), and EDTA (400 µg/ml) was added (see text). The suspension was homogenized in a Dounce homogenizer (Kontes, Vineland, NJ) and sonicated with 20 1-s blasts from a probe sonicator. The pH was adjusted to 2.0 by the addition of 2% trifluoroacetic acid, and the homogenate was stirred for 1 h at 4°C. Homogenates were centrifuged at 100,000 x g for 30 min. Supernatant was collected, and residual debris and lipid were removed by filtration (0.2 µm pore size). Low m.w. material was isolated by centrifugation through a 5,000-Da exclusion membrane (Millipore, Bedford, MA), aliquoted, and stored at -80°C. Aliquots were concentrated by Speed-Vac (Savant Instruments, Farmingdale, NY) and analyzed by HPLC.

Peptide extract from K^d

The HTR and BHA cell lines were grown and stored as described. Abs AF6-88.5.3 (IgG2a; anti-K^b) and 20-8-4S (IgG2a, anti-K^d; American Type Culture Collection, Manassas, VA) were purified from hybridoma supernatants by protein A-Sepharose, quantitated by OD₂₈₀ absorbance, and titrated by FACS analysis for staining of RMA (H-2^b) and P815 (H-2^d) cells. Extracts were prepared in batches of the following scale with all steps conducted at 4°C. Control and K^d-specific affinity columns containing 5 mg of purified Ab loaded on 2 ml of protein A-Sepharose were connected in series and equilibrated in Tris-buffered saline (TBS; 50 mM Tris and 0.9% saline) with 1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), pH 8.0. HTR or BHA cells (5 x 10⁹) were solubilized in 50 ml of TBS with 1% CHAPS, pH 8.0, containing the protease inhibitors PMSF (2 mM), iodoacetamide (100 µM), aprotinin (5 µg/ml), leupeptin (10 µg/ml), pepstatin A (10 µg/ml), and EDTA (400 µg/ml). Detergent lysates were stirred for 1 h at 4°C. Insoluble material was pelleted by centrifugation at 100,000 x g for 1 h. Supernatants were filtered (0.2 µM) and run over the protein A affinity columns twice. Columns were washed with excess TBS/1% CHAPS, pH 8.0. The control and K^d-specific columns were then separated and treated individually. Columns were washed sequentially with 5–6 column volumes of 1 M NaCl/20 mM Tris, pH 8.0, followed by 20 mM Tris, pH 8.0. Complexes were eluted from the columns with 0.2 N acetic acid and collected as 1-ml fractions. Fractions were further acidified with the addition of 100 µl of glacial acetic acid. Individual fractions were surveyed for protein content by OD₂₈₀. Fractions testing above background were pooled and boiled for 5 min. Peptides were separated from Ab and class I by centrifugation through a 5000-Da exclusion filter (Millipore). Peptide material was aliquoted and stored at -80°C.

Cytotoxicity assays

Target cells (RMA-S/K^d, HTR, BHA) were resuspended to 1–1.5 x 10⁶ cells in 0.5 ml of DMEM supplemented with 10% heat-inactivated newborn calf serum and 25 mM HEPES (assay medium). Cells were labeled with 200–250 µCi of ⁵¹Cr for 2 h at 37°C, washed three times, and plated in microtiter plates at 1 x 10⁴ cells/well in triplicate. HPLC fractions were added in 50-µl aliquots in PBS for whole cell extract or in 50-µl aliquots in PBS/10% FCS for purified K^d peptides. Synthetic peptide was added in 50 µl of assay medium. CTL were added in 50 µl of assay medium at a 10:1 E:T cell ratio for extracts and synthetic peptide or at multiple E:T cell ratios for BHA and HTR. Plates were spun for 5 min at 200 x g and incubated 4–5 h for extracts and 3 h for HTR and BHA targets at 37°C in a 10% CO₂ atmosphere. One hundred microliters of supernatant was harvested and counted. The percent specific lysis was calculated as 100% x [(experimental - spontaneous release)/(total - spontaneous release)].

Acid-stripped peptides from viable cells

BHA cells (5–6 x 10⁸) were washed three times in PBS, pH 7.3. Pellets were briefly resuspended in 10 ml of an isotonic citrate/phosphate buffer at pH 3.0 and pelleted (27, 28). The total exposure time at pH 3.0 was 4 min. Immediate quenching of the low pH with tissue culture medium resulted in a loss of viability <5% in multiple experiments. The buffer phase was collected, filtered (0.2 µm pore size), and concentrated under vacuum. The low m.w. fraction was collected by centrifugation through a 5000-Da exclusion membrane (Millipore), aliquoted, and stored at -80°C. Aliquots were concentrated by Speed-Vac and analyzed by HPLC.

Subcellular fractionation

BHA cells (10⁹) were resuspended in 10 ml of lysis buffer, 50 mM KOAc (pH 7.5), 250 mM sucrose, 6 mM MgOAc, 1 mM DTT (freshly added), 0.5 mM PMSF, 0.027 U/ml aprotinin, 1 mM EDTA, 50 mM triethanolamine (TEA; adjusted to pH 7.5 with acetic acid), and 10 µg/ml leupeptin and were disrupted by nitrogen cavitation (1,000 psi for 15 min). Debris was pelleted by centrifugation at 2000 x g for 5 min. The supernatant was collected, and membranes were pelleted by centrifugation at 100,000 x g for 1 h. Supernatant (cytosol) and membranes were handled separately, and peptide extract was generated by the TFA extraction method described for whole cell homogenates.

HPLC: standard conditions

Peptides were analyzed on a Waters HPLC system (Milford, MA) using a 4.6-mm internal diameter x 25-cm C₁₈ (Vydac, Hesperia, CA) analytical column. Gradient conditions were 0–5 min in 0–10% B, 5–35 min in 10–40% B, 35–45 min in 40–60% B, and 45–50 min in 60–0% B, where solvent A was H₂O/0.1% TFA, and solvent B was acetonitrile (ACN)/0.1% TFA at a flow rate of 1 ml/min. Peptide from whole cell homogenates (from typically 5 x 10⁸ cells) or acid-stripped cells (5 x 10⁸ cells) in H₂O/TFA or concentrated citrate/phosphate, respectively, was injected using a Waters U6K manual injection port. Peptide isolated from affinity-purified K^d (1 x 10⁹ cells) or synthetic peptide (30 µg) was injected in 10% acetic acid. One-milliliter fractions were collected. For whole cell, acid-stripped extracts, or synthetic peptide, fractions were dried completely by Speed-Vac and reconstituted in PBS, pH 7.3. For peptide material from purified K^d, fractions were concentrated by Speed-Vac to 20–30 µl and reconstituted in PBS/10% calf serum. For some fractions, the pH was adjusted to neutral by the addition of 1 N NaOH. All injections of biologically active material were preceded by injection of an equivalent volume of solvent and followed by a blank gradient run from which fractions were collected, handled identically as peptide fractions, and tested for target cell sensitization to verify the absence of retained material in the HPLC system. Data from blank runs are omitted from figures for clarity.

Size exclusion chromatography

BHA cells (3 x 10⁸) were washed three times in PBS, pH 7.3. Pellets were resuspended in lysis buffer (0.01 M Tris, pH 7.0, and 0.15 M NaCl) containing the protease inhibitor mixture previously described for purified K^d preparations. Cells were homogenized, and the membrane fraction was pelleted by centrifugation at 100,000 x g for 1 h. The cytosolic fraction was collected and applied to a Sephacryl HR S-200 (5.3 cm² x 90 cm) size exclusion column (Pharmacia, Piscataway, NJ) equilibrated with lysis buffer. Cytosolic protein was eluted with excess lysis buffer, and 5-ml fractions were collected. Marker proteins used for column calibration were ferritin (m.w. 440,000), aldolase (m.w. 158,000), albumin (m.w. 67,000), and ribonuclease A (m.w. 13,700). Individual fractions were surveyed for protein content by OD₂₈₀. Groups of 10 fractions were pooled, and peptide was isolated according to the TFA extraction protocol described for whole cell homogenates.

HPLC purification of the self peptide

Acid-soluble peptide extract from whole cell homogenates of 6 x 10⁹ HTR cells was separated by HPLC as described. Fractions 28 and 29 containing the self peptide recognized by CTL clone 14-7 were collected and pooled. Pooled fraction 28/29 peptide was rechromatographed on the same column using the same A and B buffers as those described above with a linear gradient of 0–40% B in 80 min. Self peptide eluted in fraction 55 of the second-dimension separation. This fraction was collected and again applied to the same column. Peptide was eluted with a linear gradient of 0–40% B in 80 min, where solvent A was H₂O/0.1% heptafluorobutyric acid (HFBA), and solvent B was ACN/0.1% HFBA. Self peptide eluted in fraction 60 of the third-dimension separation. This fraction was collected and again applied to the same column. Peptide was then eluted with a linear gradient of 0–40% B in 80 min, where solvent A was H₂O/5 mM NaH₂PO₄/Na₂HPO₄, pH 6.3, and solvent B was ACN/5 mM NaH₂PO₄/Na₂HPO₄, pH 6.3. This fourth-dimension separation resolved two distinct species recognized by CTL clone 14-7 eluting in fractions 41 and 45–46. These two activities were evaluated individually and rechromatographed on a pH-stable, 15-cm x 4.6-mm internal diamter, HPLC column (Jupiter 5 µ C₁₈-300A, Phenomenex, Torrance, CA). Self peptide was eluted with a linear gradient of 0–40% B in 80 min, where solvent A was H₂O/0.1% TEA, pH 8.0, and solvent B was ACN/0.1% TEA. Phosphate gradient fractions 41 and 45–46 eluted in fractions 33 and 36–37, respectively, from the fifth-dimension separation. Fractions 33 and 36–37 eluted with ACN/TEA were further analyzed by mass spectrometry.

Peptide sequence determination by tandem mass spectroscopy

HPLC fractions 33 and 37 eluted in ACN/TEA as described were evaporated to near dryness and reconstituted in 10 µl of 1% acetic acid. One-microliter aliquots were then analyzed by microcapillary HPLC electrospray-ionization tandem mass spectroscopy according to reported methods (29, 30). Briefly, peptide was eluted from a C₁₈ microcapillary column with a gradient of ACN/0.1 M acetic acid at a flow rate of 0.6 µl/min directly into the electrospray ion source of a TSQ-7000 triple quadrupole mass spectrometer (Finnigan-MAT, San Jose, CA). Abundant ion species were sequenced by interpretation of the collision-activated dissociation (CAD) spectra.

Synthetic peptides

Peptides were synthesized on a Gilson model AMS422 automated multiple peptide synthesizer (Oberlin, OH) by the solid phase method using F-moc chemistry and in situ activation. Peptides were cleaved from the resin and protecting groups by TFA with appropriate scavengers added, washed in diethyl ether, and dissolved in 10% acetic acid. Peptides were purified by HPLC and lyophilized. Lyophilized peptides were maintained at -20°C, and stock solutions were made in 100% DMSO. Peptide identity was confirmed by mass spectroscopy.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Recognition of a self peptide by an influenza HA-specific CD8⁺ CTL clone

Previous studies of the murine CTL response to the H2N2 strain influenza virus A/Japan/57 have demonstrated that the HA glycoprotein is a major target for CTL of the H-2^d haplotype (31, 32). Analysis of a panel of HA-specific CTL clones maintained in vitro identified three unique specificities restricted by the K^d class I molecule. Two distinct and partially overlapping epitopes are located in the extracellular domain and correspond to residues 204–212 and 210–219 (33, 34) (S. S. Tykodi and T. J. Braciale, unpublished observations). An additional subpopulation of K^d-restricted CTL clones recognizes an epitope from the HA transmembrane domain. Clones specific for this epitope were shown to recognize H-2^d target cells treated with a 23-aa synthetic peptide corresponding to the HA transmembrane sequence 523–545 (single-letter amino acid code, VYQILAIYATVAGSLSLAIMMAG) (20, 21). Two peptides with a truncation from either the amino (531–545) or the carboxyl (523–537) end of the proband 23-residue peptide were still able to sensitize target cells for recognition by three transmembrane domain-specific CTL clones, indicating an apparent requirement for the overlapping residues (531–537, ATVAGSL) (35).

The sequence motif for the K^d molecule has been used to predict naturally processed peptide epitopes recognized by CTL (36, 37, 38). The HA_523–545 transmembrane domain contains a single K^d motif sequence (defined as a Y or F at position 2 (P2) relative to the amino terminus and I, L, or V at the C-terminus in sequences 9 or 10 aa in length) (14, 39, 40, 41) at residues 529–537 (IYATVAGSL). This sequence contains the 531–537 core sequence shown to be essential for CTL recognition of the longer transmembrane domain proband peptide. A synthetic peptide corresponding to the HA residues 529–537 efficiently sensitized H-2^d P815 target cells for recognition by four independently isolated CTL clones specific for the HA transmembrane domain (data not shown; see Fig. 10). Consistent with previous data defining the immunodominance of the 523–545-specific response, polyclonal, influenza-specific CTL cultures raised by in vitro stimulation of splenocytes from A/Japan/57-immune BALB/c mice are potent cytotoxic effectors when tested for recognition of P815 target cells pulsed with the 529–537 peptide (V. Foster and T. J. Braciale, unpublished observation).

View larger version (28K): [in this window] [in a new window]   FIGURE 10. Recognition of synthetic nonamer and decamer peptides related to the proband (E)NYAYPG(I/L)(I/L) peptide. All possible sequence permutations of the core peptide sequence (E)NYAYPG(I/L)(I/L) defined by mass spectroscopy-based sequencing of the cross-reactive self peptide purified from HTR cells were synthesized. Stock solutions of synthetic peptides were diluted and tested for the capacity to sensitize 51Cr-labeled RMA-S/Kd target cells for lysis by CTL clone 14-7 in a 4- to 5-h assay. Additional assay parameters are described in Fig. 1.

Surprisingly, HPLC fractions from whole cell homogenates of the BHA cell line contained two distinct species capable of sensitizing RMA-S/K^d target cells for recognition by CTL clone 14-7. The transmembrane domain-specific clone (20, 21) detected peptide eluting in HPLC fractions 23–25 as well as fractions 28–29 (Fig. 1A). Synthetic HA_529–537 peptide chromatographed under identical gradient conditions yielded peptide in HPLC fractions 24–25 (Fig. 1B) identifying the first peptide eluting from the BHA extract as the naturally processed K^d motif peptide 529–537. To determine whether recognition of HPLC fractions 28–29 by clone 14-7 was linked to the expression of HA protein in BHA, peptide extract was prepared from the control HTR cell line for comparison. The peptide extract from HTR was separated by HPLC, and fractions were tested for their capacity to sensitize target cells for recognition by the 14-7 CTL clone. While peptide eluting in fractions 23–25 corresponding to the HA_529–537 epitope was predictably absent in the HTR-derived peptides, fractions 28–29 were again recognized by clone 14-7 (Fig. 1C). Rather than representing an HA-derived structure related to the 529–537 peptide, the activity in HPLC fractions 28–29 must come from a source endogenous to the HTR cell line. In light of the isolation of this activity as an acid-extractable, low m.w. moiety with an elution profile from the C₁₈ HPLC column matrix similar to that of MHC-derived peptides, it was probable the activity present in HPLC fractions 28–29 represented an unexpected cross-reactivity by clone 14-7 for an endogenous peptide. The HPLC fractions of extract from whole cell homogenates of BHA recognized by clone 14-7 were titrated. Fractions 28–29 were significantly more potent in sensitizing RMA-S/K^d target cells for recognition by clone 14-7 than were fractions 23–24 containing the HA_529–537 peptide (Fig. 1D).

View larger version (16K): [in this window] [in a new window]   FIGURE 1. CTL clone 14-7 recognizes both HA and self peptide in HPLC fractions of peptide from whole cell homogenates of the BHA cell line. Low m.w., acid-soluble peptide isolated from whole cell homogenates of 6 x 108 BHA (A) or HTR (C) cells or 30 µg of HA529–537 (B) was fractionated by HPLC, and fractions were tested for the capacity to sensitize 51Cr-labeled RMA-S/Kd target cells for lysis by CTL clone 14-7 in a 4- to 5-h assay. Fractions were resuspended in PBS, pH 7.3, and divided in half for A and C. One-half of the biologically active fractions from A were serially diluted in PBS and tested for target cell sensitization (D). Peptide fractions in B were diluted 4 x 108 in PBS. All assays were performed in triplicate with clone 14-7 added at a 10:1 E:T cell ratio.

View larger version (16K): [in this window] [in a new window]   FIGURE 2. CTL clone 14-7 recognition of self peptide is Kd restricted. Low m.w., acid-soluble peptide was isolated from whole cell homogenates of 6 x 108 BHA cells, and HPLC fractions were tested for the capacity to sensitize 51Cr-labeled untransfected RMA-S or RMA-S/Kd target cells for lysis by clone 14-7. Individual fractions were divided in half and assayed on both target cells. Assay parameters are described in Fig. 1.

View larger version (28K): [in this window] [in a new window]   FIGURE 5. Presence of the cross-reactive self peptide is independent of MHC class I or ß2m. Low m.w., acid-soluble peptide was isolated from whole cell homogenates of 5–6 x 108 nucleated splenocytes (A–D) or cells from in vitro culture (E and F). Peptide was separated by HPLC, and fractions were tested in a cytotoxicity assay as described in Fig. 1. The source of peptide is indicated: A, BALB/c (H-2d); B, C57BL/6 (H-2b); C, C3H (H-2k); D, ß2m knockout (ß2m knockout backcrossed into the BALB/c background, H-2d); E, R1.E (ß2-m-, H-2k); and F, S49.35 (classical MHC class I deficient, H-2d).

As shown above, CTL recognition of a high titer self peptide in HPLC fractions 28–29 of low m.w. peptide recovered from whole cell homogenates of multiple sources, including in vitro cultured tumor lines and freshly isolated splenocytes, was K^d restricted. Further, the self peptide may well contain a typical K^d sequence motif. The CTL clone 14-7 had not previously been noted to have an autoreactive phenotype (20, 21, 32, 35). Nevertheless, a K^d-associated self peptide seemed the most likely source of the cross-reactivity present in whole cell homogenates and recognized by the 14-7 clone. To test this hypothesis directly, the K^d molecule was immunoaffinity purified from detergent lysates of BHA and HTR cell lines. Acid-extractable peptide material was separated from denatured class I heavy chain and ß₂m and concentrated. The K^d-derived peptide mixture was separated by HPLC, and fractions were tested for their ability to sensitize target cells for recognition by clone 14-7. As shown in Fig. 3, clone 14-7 recognized K^d-derived peptide from the BHA cell line eluting in HPLC fractions 24–25, which was indistinguishable from synthetic HA_523–545 (see Fig. 1B). The total heterogeneous peptide material isolated from affinity-purified K^d from the BHA and HTR cell lines failed to yield peptide detectable by the 14-7 CTL clone in HPLC fraction 28 or 29 (Fig. 3 and data not shown).

View larger version (16K): [in this window] [in a new window]   FIGURE 3. HA epitope, but not cross-reactive self peptide, is recovered from immunoaffinity purified Kd molecules from the BHA cell line. The Kd molecule was affinity purified from a detergent lysate of the BHA cell line with the mAb 20-8-4S (IgG2a). A mock preparation using mAb AF6-88.5.3 (IgG2a, anti-Kb) was used as a control. Peptide from purified Kd representing 1 x 109 cells was fractionated by HPLC. Individual fractions were concentrated, resuspended in PBS/10% FCS, and tested for the capacity to sensitize 51Cr-labeled RMA-S/Kd target cells for lysis by CTL clone 14-7. Assay parameters are described in Fig. 1.

View larger version (14K): [in this window] [in a new window]   FIGURE 4. Recognition of the HTR and BHA cell lines by a panel of HA transmembrane domain-specific CTL clones. HTR (A and B) or BHA (C) cells were 51Cr labeled and used as targets in a 3-h cytotoxicity assay. CTL clones 14-7, A4, and 19-10 are specific for the Kd-restricted HA529–537 epitope. Clone 19-11 is specific for the influenza NP147–155 epitope. HTR targets in B are pulsed with synthetic peptide HA529–537 (100 nM) or NP147–155 (1 nM).

Although the haplotype-independent phenotype of the self peptide recognized by clone 14-7 does not formally exclude the possibility of a promiscuous peptide binding to multiple class I molecules, the current understanding of allele-specific sequence motifs of class I molecules makes this scenario less likely. An alternative possibility considered for the source of self peptide was the promiscuous binding of a peptide derived from one of the nonpolymorphic MHC class Ib allelic products. Members of this group of class I-related proteins are known to bind peptides and offered a possible explanation for the high titers of self peptide and the haplotype independence previously described (48, 49, 50, 51). To investigate a requirement for properly assembled cell surface MHC to detect the HPLC fraction 28–29 self peptide, peptide extract was made from whole cell homogenates of splenocytes from mutant mice with a disrupted ß₂m gene (ß₂ knockout) that are formally H-2^d haplotype. Cells from these mice fail to express stable MHC class I heterodimers on the cell surface because of a targeted disruption of the ß₂m gene (52, 53). Nevertheless, peptide extract derived form splenocytes of these mice contained a self peptide in HPLC fractions 28–29 detected by clone 14-7 (Fig. 5D). When the fractions containing self peptide were titrated, splenocytes from ß₂ knockout mice contained a similar amount of peptide in fractions 28–29 as the other splenocyte-derived samples (data not shown). Comparable results were obtained with peptide extracts of whole cell lysates from the ß₂m-deficient, H-2^k haplotype, R1.E cell line (25) (Fig. 5E) and from the S49.35 cell line (Fig. 5F), which does not express classical MHC class I molecules (24). These results imply that the presence of the self peptide detected in HPLC fractions 28–29 of peptide extracts from whole cell lysates does not depend on the presence of properly conformed conventional MHC class I molecules.

The lack of an association of the presence of the self peptide detected by clone 14-7 and MHC expression was unexpected, as previous reports have demonstrated a requirement for the presence of the restricting MHC allele to biochemically isolate peptide epitopes recognized by CTL (39, 54, 55). We reconsidered the issue of the association of the self peptide recognized by clone 14-7 with a class I-related molecule by using a technique for isolating class I-associated peptides from the surface of viable cells. Peptides were isolated from the surface of viable BHA cells by treating the cells with an isotonic buffer at pH 3.0, thereby denaturing class I molecules and eluting the associated peptides into the buffer phase (27, 28). Separation of low m.w. material and HPLC fractionation resulted in the efficient recovery and detection of the HA_529–537 peptide eluting in fractions 24–25 (Fig. 6). However, there was no recognition of fractions 28–29 by clone 14-7, indicating that the self peptide readily isolated from whole cell lysates is not present at the cell surface.

View larger version (13K): [in this window] [in a new window]   FIGURE 6. Acid stripping viable BHA cells recovers only HA peptide and not the cross-reactive self peptide. BHA cells (5–6 x 108) in culture were washed three times in PBS, pH 7.3, and briefly exposed to an isotonic citrate/phosphate buffer at pH 3.0. Supernatants were collected and concentrated, and low m.w. material was isolated and fractionated by HPLC. Fractions were resuspended in PBS, pH 7.3, divided in half, and tested in a cytotoxicity assay as described in Fig. 1.

View larger version (17K): [in this window] [in a new window]   FIGURE 7. The HA and cross-reactive self peptide differentially segregate to membrane and cytosol fractions from BHA cells. Membrane (A) and cytosol (B) components from 1 x 109 BHA cells were separated. Low m.w., acid-soluble peptide was isolated from the two components separately according to the TFA extraction method established for whole cell homogenates. Peptides were separated by HPLC. Fractions were resuspended in PBS, pH 7.3, divided in half, and tested in a cytotoxicity assay as described in Fig. 1. One-half of the biologically active fractions from A and B (24 and 29) was serially diluted in PBS and tested for target cell sensitization (C, fraction 24; D, fraction 29) under similar assay conditions.

The self peptide recognized by clone 14-7 was localized to the cell cytoplasm, and the expression of this self peptide in the cytosol did not depend on the presence of MHC class I molecules. It was of interest to determine whether the self peptide was present in the cytosol as a free peptide or bound to a cytosolic carrier protein. Lysates of BHA cells were again separated into a membrane and a cytosol component. The cytosolic fraction of BHA cells was further fractionated by size exclusion chromatography at neutral pH. Fig. 8A shows the absorbance at 280 nm of the fractionated cytoplasm and the elution profile of molecular mass standards used to calibrate the separation column. Cytosolic proteins larger than the exclusion limit of the column (200 kDa) eluted between column fractions 178 and 227, while free peptides less than 5 kDa were estimated to elute between fractions 425 and 478. To detect the presence of the self peptide in the column-separated cytosol, fractions corresponding to six distinct molecular mass ranges spanning from <5 to >200 kDa were pooled to yield six fractions of 50 ml each. Low m.w. acid-soluble peptide was then isolated from each pooled fraction. The peptides associated with each pool were subjected to HPLC separation and tested for target cell sensitization and recognition by clone 14-7. As Fig. 8B demonstrates, the bulk of the self peptide activity was contained in eluate fractions 228–277, corresponding to a mass between 40–150 kDa. Little or no peptide activity was detected in the mass range of free peptides (fractions 426–478).

View larger version (21K): [in this window] [in a new window]   FIGURE 8. Size exclusion chromatography of BHA cytosol at neutral pH. Membrane and cytosol components from 3 x 108 BHA cells were separated. BHA cytosolic protein in TBS at pH 7.0 was applied to a Sephacryl S-200 column and eluted with excess TBS. Five-milliliter fractions were collected, and protein content was quantitated by OD280 (A). Retention times for marker proteins are indicated in the figure and include ferritin (m.w. 440,000), aldolase (m.w. 158,000), albumin (m.w. 67,000), and ribonuclease A (m.w. 13,700). Eluate was consolidated into six 10-fraction groups (50 ml each). Each group was acidified to pH 2.0 by addition of TFA. Low m.w. material was collected and fractionated by HPLC. Fraction 29 from each group was tested in a cytotoxicity assay as described in Fig. 1 (B).

To identify this putative K^d-binding self peptide, we conducted multidimensional HPLC separation of HTR extract using various elution buffer modifiers to enrich for the active material. Fig. 9 shows HPLC profiles of the UV absorbance (A214 nm), elution gradients, and clone 14-7 recognition of peptide-loaded target cells for each HPLC separation step. Briefly, the biologically active material in fractions 28–29 of our standard acetonitrile/TFA gradient (Fig. 9A) was rechromatographed, employing a shallow gradient with the same acetonitrile/TFA mobile phase (Fig. 9B). Further purification was achieved by HPLC separation of the active material in Fig. 9B using HFBA as the mobile phase modifier (Fig. 9C). The active material in Fig. 9C was further analyzed using inorganic phosphate (pH 6.3) as the modifier. Two peaks of biological activity were resolved by this separation step in fractions 41 and 45–46 (Fig. 9D). Each of these activity peaks was then subjected to further HPLC separation using TEA at pH 8.0 as the modifier. The material in phosphate peak fraction 41 eluted as a single peak of biological activity at fraction 33 in the acetonitrile/TEA gradient (Fig. 9E). The material in phosphate peak 45 was likewise resolved into one dominant activity peak in fractions 36–37 after acetonitrile/TEA separation (Fig. 9F).

View larger version (40K): [in this window] [in a new window]   FIGURE 9. HPLC purification of cross-reactive self peptide from HTR cells. Low m.w., acid-soluble peptide was isolated from whole cell homogenates of a total of 5 x 109 HTR cells. Peptide was separated by HPLC under five different gradient conditions, and fractions were tested in a cytotoxicity assay as described in Fig. 1. Fractions with biological activity were collected and rechromatographed. Gradient conditions employed elute peptide from an aqueous solution with the following: A, ACN/0.1% TFA by our standard separation gradient; B, ACN/0.1% TFA shallow gradient; and C, ACN/0.1% HFBA; D, ACN/5 mM NaH2PO4/Na2HPO4 at pH 6.3. The two separate activities from D (fractions 41 and 45–46) were collected and rechromatographed separately with ACN/0.1% TEA at pH 8.0 (E, fraction 41; F, fractions 45–46).

In contrast to other abundant peptides present in HPLC fractions 33 and 37 of the acetonitrile/TEA gradient, only the two related peptides with the sequence of (E)NYAYPGV (I/L)(I/L) contained a K^d sequence motif. We synthesized a panel of peptides with all possible sequence permutations of isoleucine and/or leucine at the C-terminal and subterminal positions of both the nonamer and decamer peptides and tested the capacities of these peptides to sensitize RMA-S/K^d targets for recognition by clone 14-7. As Fig. 10 demonstrates, all the nonamer and decamer peptides tested were recognized to varying degrees. Four of the sequences tested were at least equipotent compared with the control HA_529–537 peptide and able to sensitize RMA-S/K^d target cells for recognition by clone 14-7 at doses as low as 10 pM. The synthetic peptides corresponding to the other abundant peptide sequences in HPLC fractions 33 and 37 identified by CAD failed to sensitize target cells for recognition by clone 14-7 at any dose tested (data not shown).

A search of the protein and nucleic acid databases revealed that the sequence ENYAYPGVLL precisely matches residues 179–188 of the mammalian mitochondrial aconitase precursor protein. This enzyme, which is highly conserved in mammalian species (>90% protein sequence homology among bovine, porcine, and human enzymes), is encoded by a nuclear gene, synthesized on free ribosomes in the cytosol, and then targeted post-translationally to the mitochondrial matrix (56, 57). Although the amino acid sequence of the murine aconitase protein has not been completely determined, the derived sequence of partial cDNAs encoding the portion of the murine enzyme spanning residues 179–188 is identical with that of the (E)NYAYPGVLL self peptide recognized by clone 14-7. We evaluated the panel of nonamer and decamer synthetic peptides containing all possible sequence permutations of isoleucine and/or leucine at the C-terminal and subterminal positions and determined that only the decamer ENYAYPGVLL and the nonamer NYAYPGVLL peptides had the same HPLC elution profile in the acetonitrile/TEA gradient as the biologically active material in HPLC fractions 33 and 37, respectively (data not shown). Further, ion species corresponding to the molecular masses of the two peptides we have identified (1101 and 1039) were present in total cell acid extract from freshly isolated BALB/c mouse splenocytes. Sequences for these mass species determined by CAD spectra were identical with the peptides detected in extracts from the HTR cells (data not shown). These findings confirm that the two peptides containing the sequence (E)NYAYPGVLL fully account for the self peptide recognized in a cross-reactive fashion by clone 14-7.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
In this report we have identified two peptides recognized by a murine influenza HA-specific, K^d-restricted, CD8⁺ CTL clone. The specificity of the CTL clone 14-7 had previously been mapped to the transmembrane domain of the HA glycoprotein. The naturally processed, K^d-associated ligand recognized by this clone was identified as the K^d motif peptide corresponding to residues 529–537 (IYATVAGSL) of the HA transmembrane domain. This peptide is predictably associated with the restricting K^d molecule intracellularly and can be recovered from purified K^d molecules. Surprisingly, the CTL clone 14-7 also recognizes an abundant self peptide derived from the highly conserved, ubiquitous, murine mitochondrial aconitase enzyme. This self peptide, corresponding to the aconitase precursor protein residues 179–188 ((E)NYAYPGVLL) is detectable in total cell extracts of lymphoreticular cells from a variety of sources, including murine cell lines and freshly isolated splenocytes. Recognition of the aconitase peptide was shown to be K^d restricted and clone 14-7 specific. Of particular interest is the finding that this self peptide is not associated with its restricting MHC class I molecule on the cell surface. Rather, it is sequestered in the cytoplasm and probably associated with a carrier protein.

Several lines of evidence suggest that this mitochondrial aconitase peptide, although requiring association with the K^d molecule for CTL recognition in vitro (Fig. 2), is not associated with this or other MHC class I allele under physiologic conditions in intact cells. First, intact target cells containing this self peptide are not recognized in a cross-reactive manner by the HA_529–537-specific CTL clone 14-7 (Fig. 4A). Second, the presence of this self peptide in extract from whole cell homogenates is not dependent on the expression of a particular MHC class I allele or ß₂m (Fig. 5). Third, while the naturally processed K^d-associated HA_529–537 peptide can be released from the surface of intact cells by low pH acid stripping and is localized to the MHC class I-containing membrane fraction of disrupted cells in neutral lysis buffer, the aconitase peptide cannot be stripped from the cell surface and is localized exclusively to the cell cytoplasm (Figs. 6 and 7).

There is a large body of compelling evidence suggesting that self and foreign peptides recognized by CD8⁺ CTLs are primarily associated with specific MHC class I molecules inside intact cells (39, 54, 55). This intimate association with MHC class I molecules is believed to protect peptides from rapid proteolysis by ubiquitous proteases present throughout the cell. As noted above, we were unable to demonstrate that the presence of the 9/10-aa aconitase peptide inside the cell depended on its association with MHC class I molecules. The results reported here suggest that this aconitase-derived peptide is neither free within the cell cytoplasm nor bound to MHC class I molecules in the cells. Rather this self peptide exists preformed in the cell cytoplasm tightly associated with one or more cytosolic carrier moieties (Fig. 8). The carrier protein associated with the aconitase peptide may serve a protective function, allowing for the intracellular accumulation of this peptide.

We considered the possibility that this abundant self peptide is generated by proteolytic cleavage of its precursor protein through the action of cellular proteases liberated during TFA acid extraction. Although disruption of cells at a pH value as low as 2.0 with a strong denaturing reagent such as TFA should inactivate most intracellular proteases, fragmentation of the aconitase enzyme by a TFA-resistant cellular protease during cell disruption with generation of the aconitase peptide cannot be formally excluded. However, we were unable to demonstrate any significant difference in either the abundance of this self peptide or the relative proportion of the nonamer and decamer peptides when cells were extracted in the presence or the absence of a mixture of protease inhibitors (data not shown). Likewise, prolonged incubation of TFA-treated cell extracts before separation of the low m.w. material failed to show any increase in the relative abundance of the aconitase peptide (data not shown). Furthermore, the aconitase peptide was not found as free peptide in whole cell homogenates. Rather it was colocalized almost exclusively with cytosolic proteins of a particular mass range (40 and 150 kDa; Fig. 8). These observations favor the view that the 9/10-aa aconitase peptides exist in the cell before cell disruption for peptide extraction.

These findings raise the question: how is this abundant self peptide excluded from the class I pathway and not presented on the cell surface? Mitochondrial proteins are encoded by both mitochondrial and nuclear (chromosomal) genes. Several mitochondria-encoded proteins contain amino-terminal N-formylated peptide epitopes recognized by CD8⁺ CTLs restricted by nonclassical MHC class Ib molecules (51). The aconitase enzyme, however, is encoded by a nuclear gene and is likely post-translationally targeted to the mitochondria and translocated to the mitochondrial matrix after synthesis on free ribosomes in the cytoplasm (56). Therefore, a cytosolic precursor of the active form of the enzyme in the mitochondrial matrix is not only accessible to chaperone proteins required for mitochondrial targeting and translocation, but is potentially available for partial proteolysis by cytosolic proteases. Nonetheless, the substrate for proteases is not necessarily the cytosolic precursor, but may be the mature protein exported to the cytosol to be degraded (58). If, as our data suggest, this self peptide pre-exists in the cell cytoplasm of intact cells as a 9/10-aa fragment capable of binding MHC class I molecule, then its absence on the cell surface is not due to a defect in proteolytic processing of the nascent or mature mitochondrial aconitase protein. Rather, the failure of this peptide to be presented may reflect inefficient transport of the peptide from the cytoplasm to the ER by the TAP1/TAP2-associated transporter complex (59). Similarly, this peptide may be transported into the ER but be unable to form a stable complex with the nascent MHC class I molecule and its associated ER chaperone proteins in the ER despite its capacity to efficiently bind preassembled, properly conformed, class I molecules at the cell surface (60).

Because the aconitase peptide identified in this report coelutes with cytosolic proteins of a mass comparable to that of cytosolic hsp chaperones, it is tempting to speculate that the association of this self peptide with a chaperone of the hsp70/90 stress protein family accounts for both the protection of the mitochondrial peptide from proteolytic degradation and its sequestration in the cytoplasm. There is precedent for the association of antigenic peptides inside intact cells with molecules other than MHC proteins, notably, hsp family chaperones (8, 9). The hsp are among the most abundant proteins in the cell cytoplasm. Fully processed antigenic peptides capable of binding MHC class I molecules have been demonstrated to be bound to the hsp70 chaperone protein isolated from cells expressing a foreign Ag (9). We favor the model that the lack of presentation of this abundant self peptide by the K^d molecule may reflect a targeting failure of the aconitase peptide/chaperone complex to associate with the TAP1/TAP2 transporter or an inability of the hydrophobic aconitase peptide to efficiently dissociate from its cytosolic chaperone under the neutral pH conditions of the cell cytoplasm before transport by the TAP complex. Consistent with the later view, available data suggest that the aconitase peptide can only be released from the cytosolic chaperone under conditions of low pH (R. Fan and T. J. Braciale, unpublished observation). Further studies are necessary to characterize the cytosolic protein to which the aconitase 179–188 peptide is bound and the nature of the interaction between peptide and chaperone.

At present, we do not know whether peptide sequestration in the cytoplasm is a property exhibited by most cellular proteins that undergo proteolytic degradation in the cytoplasm or is a characteristic of a specific class of proteins, for example, proteins targeted to intracellular compartments that may require chaperone assisted translocation (61). Similarly, we do not as yet know the frequency with which foreign Ag-specific CD8⁺ T lymphocytes will cross-react with sequestered self peptide. Our preliminary data suggest that peptide sequestration is not limited to the one example described in this report. In screening a panel of 15 influenza-specific CTL clones, we found a second CTL clone that recognized a unique collection of self peptides. The CTL clone RK21 is K^d restricted and specific for the influenza HA peptide corresponding to residues 210–219. This clone does not recognize the aconitase peptide 179–188, but does cross-react to multiple HPLC fractions of self peptide from splenocyte homogenates. The peptides recognized cross-reactively by RK21 have a phenotype similar to the aconitase peptide that includes clone-specific, K^d-restricted recognition as well as expression in splenocytes independent of haplotype or ß₂m expression. The sequence identity of these peptides remains to be established. We also note the report by Rötzschke et al. documenting a self peptide recognized by an alloreactive CTL clone that is MHC unrestricted in expression, is even detectable in the human Jurkat cell line, and is not associated with the purified restricting class I MHC molecule (62). This peptide may represent another example of a sequestered self peptide. Taken together, accumulating evidence suggests that the sequestered aconitase self peptide described in this report is likely to be one example from a pool of heterogeneous intracellular peptides capable of binding to class I molecules and sensitizing target cells for recognition by CTL.

Although it is unknown to what extent the entirety of self protein transcribed in a cell can be recruited into the class I pathway to generate peptide for MHC assembly, there may be good reason why a limit on the total diversity of self peptide presented by MHC is advantageous. For example, it appears that both positive and negative selection of T lymphocyte precursors in the thymus is peptide dependent (10). The expression of unique self peptide/MHC complexes may be optimized for T cell repertoire diversity. Our data would predict that expression of the aconitase self peptide (or self peptide recognized by clone RK21) in the thymus would lead to the deletion of these specificities and their absence in the periphery. Sequestration of self peptides in the cell cytoplasm might thereby maximize the repertoire of CD8⁺ T lymphocytes directed to foreign Ags.

Despite theoretical benefit, peptide sequestration would also seem to pose hazards. It can be imagined that microbial Ag may be similarly sequestered and go undetected by CTL. In addition, there is now ample evidence that complexes between antigenic peptides and cytosolic or ER resident hsp are potent inducers of CD8⁺ T lymphocyte responses. These Ag-hsp complexes may use specific cell surface receptors to internalize the peptide-chaperone complex into a vesicular (endosomal) compartment where antigenic peptides can charge MHC class I molecules (2, 6). Our results suggest that the Ag receptors on activated CD8⁺ effector CTL directed to foreign Ags are capable of cross-reactive recognition of normally sequestered self peptides. The possibility arises that self peptides bound to hsp-like chaperones when released from virus-infected cells killed by specific effector T lymphocytes could sensitize uninfected bystander cells for cross-reactive recognition and destruction by virus-specific effector CTLs.

	Acknowledgments

 
We thank Julie Burns for secretarial support, and Barbara Small and Vicki Forster for technical support.

	Footnotes

 
¹ This work was supported by U.S. Public Health Service Grants AI15608, AI28317, and HL33391 (to T.J.B.) and by the Medical Scientist Training Program at Washington University, St. Louis, MO (to S.S.T.).

² R.F. and S.S.T. contributed equally to this work.

³ Address correspondence and reprint requests to Dr. Thomas J. Braciale, Beirne Carter Center for Immunology Research, Lane Road, MR-4 Building, Room 4012, University of Virginia, Charlottesville, VA 22908. E-mail address:

⁴ Abbreviations used in this paper: ß₂m, ß₂-microglobulin; ER, endoplasmic reticulum; hsp, heat shock protein; HA, hemagglutinin; TEA, triethanolamine; ACN, acetonitrile; HFBA, heptafluorobutyric acid; CAD, collision-activated dissociation; NP, nucleoprotein; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate.

Received for publication October 1, 1999. Accepted for publication November 15, 1999.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Pamer, E., P. Cresswell. 1998. Mechanisms of MHC class I-restricted antigen processing. Annu. Rev. Immunol. 16:323.[Medline]
2. Srivastava, P. K., A. Menoret, S. Basu, R. J. Binder, K. L. McQuade. 1998. Heat shock proteins come of age: primitive functions acquire new roles in an adaptive world. Immunity 8:657.[Medline]
3. Li, Z., P. K. Srivastava. 1993. Tumor rejection antigen gp96/grp94 is an ATPase: implications for protein folding and antigen presentation. EMBO J. 12:3143.[Medline]
4. Udono, H., P. K. Srivastava. 1993. Heat shock protein 70-associated peptides elicit specific cancer immunity. J. Exp. Med. 178:1391.[Abstract/Free Full Text]
5. Udono, H., P. K. Srivastava. 1994. Comparison of tumor-specific immunogenicities of stress-induced proteins gp96, hsp90, and hsp70. J. Immunol. 152:5398.[Abstract]
6. Schild, H., D. Arnold-Schild, E. Lammert, H.-G. Rammensee. 1999. Stress proteins and immunity mediated by cytotoxic T lymphocytes. Curr. Opin. Immunol. 11:109.[Medline]
7. Suto, R., P. K. Srivastava. 1995. A mechanism for the specific immunogenicity of heat shock protein-chaperoned peptides. Science 269:1585.[Abstract/Free Full Text]
8. Nieland, T. J. F., M. C. A. A. Tan, M. Monnee-van Muijen, F. Koning, A. M. Kruisbeek, G. M. van Bleek. 1996. Isolation of an immunodominant viral peptide that is endogenously bound to the stress protein GP96/GRP94. Proc. Natl. Acad. Sci. USA 93:6135.[Abstract/Free Full Text]
9. Breloer, M., T. Marti, B. Fleischer, A. von Bonin. 1998. Isolation of processed, H-2Kb-binding ovalbumin-derived peptides associated with the stress proteins HSP70 and GP96. Eur. J. Immunol. 28:1016.[Medline]
10. Jameson, S. C., M. J. Bevan. 1998. T-cell selection. Curr. Opin. Immunol. 10:214.[Medline]
11. Tanchot, C., F. A. Lemonnier, B. Pérarnau, A. A. Freitas, B. Rocha. 1997. Differential requirements for survival and proliferation of CD8 naive or memory T cells. Science 276:2057.[Abstract/Free Full Text]
12. Malnati, M. S., M. Peruzzi, K. C. Parker, W. E. Biddison, E. Ciccone, A. Moretta, E. O. Long. 1995. Peptide specificity in the recognition of MHC class I by natural killer cell clones. Science 267:1016.[Abstract/Free Full Text]
13. Zappacosta, F., F. Borrego, A. G. Brooks, K. C. Parker, J. E. Coligan. 1997. Peptides isolated from HLA-Cw*0304 confer different degrees of protection from natural killer cell-mediated lysis. Proc. Natl. Acad. Sci. USA 94:6313.[Abstract/Free Full Text]
14. Falk, K., O. Rotzschke, S. Stevanovic, G. Jung, H. G. Rammensee. 1991. Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC molecules. Nature 351:290.[Medline]
15. Ruppert, J., J. Sidney, E. Celis, R. T. Kubo, H. M. Grey, A. Sette. 1993. Prominent role of secondary anchor residues in peptide binding to HLA-A2.1 molecules. Cell 74:929.[Medline]